福建平潭近海赤潮预警模型研究*

doi:10.11978/2021156

热带海洋学报 ›› 2022, Vol. 41 ›› Issue (4): 172-180.doi: 10.11978/2021156CSTR: 32234.14.2021156

• 海洋环境保护 • 上一篇

福建平潭近海赤潮预警模型研究^*

苏金洙^1,²(), 邹嘉澍¹, 苏玉萍^1,²(), 张明峰³, 翁蓁洲⁴, 杨小强⁴

1.福建师范大学环境科学与工程学院, 福建福州 350007
2.福建省河湖健康研究中心, 福建福州 350007
3.福建师范大学地理科学学院, 福建福州 350007
4.福州市海洋与渔业技术中心, 福建福州 350007

收稿日期:2021-11-14 修回日期:2022-02-28 出版日期:2022-07-10 发布日期:2022-02-25
通讯作者: 苏玉萍
作者简介:苏金洙(1994—), 女, 福建泉州市人, 博士研究生, 从事水污染生态修复与预警研究。email: jnzusu@126.com
基金资助:
国家重点研发计划项目(2016YFE0202100);福建省高校产学合作项目(SC-292);福建省高校产学合作项目(21NB000922)

Study on the early warning model of red tide in the offshore area of Pingtan, Fujian province

SU Jinzhu^1,²(), ZOU Jiashu¹, SU Yuping^1,²(), ZHANG Mingfeng³, WENG Zhenzhou⁴, Yang Xiaoqiang⁴

1. Environmental Science and Engineering College, Fujian Normal University, Fuzhou 350007, China
2. Fujian Province Research Centre for River and Lake Health Assessment, Fuzhou 350007, China
3. Geographical Sciences College, Fujian Normal University, Fuzhou 350007, China
4. Marine and Fisheries Technology Centre, Fuzhou 350007, China

Received:2021-11-14 Revised:2022-02-28 Online:2022-07-10 Published:2022-02-25
Contact: SU Yuping
Supported by:
National Key R&D Program Funded Projects(2016YFE0202100);Industry-University Cooperation Project of Fujian Province, China(SC-292);Industry-University Cooperation Project of Fujian Province, China(21NB000922)

摘要/Abstract

摘要：

本文分析了福建省平潭近海海域2013—2019年水文、水质及气象数据的主成分结果, 筛选出5个气象因子和4个水质因子作为输入指标, 以藻密度为输出指标, 分别演算了KNN (K-nearest neighbor)、RF (random forest)、GBRT (gradient-boosted regression Trees)以及Bagging (bootstrap aggregating)4种赤潮预警回归模型。对2013—2019年的802 组海洋监测数据归一化处理后, 随机选取80%的数据作为模型的训练样本, 剩余的20%作为模型验证数据。其中, 以风速、气温、海平面气压、叶绿素a浓度组合为输入指标时, KNN回归模型演算结果的精度较高(R²=0.624, RMSE=0.821μg·L^-1, MAE=0.836μg·L^-1)。在没有叶绿素a浓度监测指标的海域, 构建了以叶绿素a浓度为输出指标, 气温、日照、风速、AOI(apparent oxygen increase)组合为输入指标的BP神经网络赤潮模型, 该模型也具有较好的预警精度(R²=0.651, RMSE=0.062μg·L^-1, MAE=0.033μg·L^-1)。本研究结果可为平潭海域的赤潮预警研究提供参考。

关键词: 叶绿素a浓度, 藻类密度, 赤潮, 预警模型, 平潭海域

Abstract:

We analyzed the principal components of hydrology, water quality, and meteorological data in Pingtan, Fujian province from 2013 to 2019. We selected 5 meteorological factors and 4 water quality factors. Our study establishes four early-warning model, KNN (K-nearest neighbor), RF (random forest), GBRT (gradient-boosted regression trees), Bagging (bootstrap aggregating) with meteorological factors and water quality factors as input indicators, and algal cell density as output indicators. After normalizing the 802 sets of marine monitoring data from 2013 to 2019, 80% of the data were randomly selected as the model training samples, and the remaining 20% were used as data of model verification. When temperature, wind speed, sea level pressure, and chlorophyll a are used as input parameters, the calculation result of KNN regression model is more accurate (R²=0.624, RMSE=0.821 μg·L^-1, MAE=0.836 μg·L^-1). In the sea area without chlorophyll a monitoring index, a BP neural network early-warning model with chlorophyll a concentration as the output index and temperature, sunshine, wind speed and AOI as input parameters was established, which has better warning accuracy (R²=0.651, RMSE=0.062 μg·L^-1, MAE=0.033 μg·L^-1). Our results can provide a reference for the red tide early warning research in the Pingtan coastal area.

Key words: chlorophyll a concentration, algal cell density, red tide, early-warning model, Pingtan coastal area

中图分类号:

P762.33

苏金洙, 邹嘉澍, 苏玉萍, 张明峰, 翁蓁洲, 杨小强. 福建平潭近海赤潮预警模型研究^*[J]. 热带海洋学报, 2022, 41(4): 172-180.

SU Jinzhu, ZOU Jiashu, SU Yuping, ZHANG Mingfeng, WENG Zhenzhou, Yang Xiaoqiang. Study on the early warning model of red tide in the offshore area of Pingtan, Fujian province[J]. Journal of Tropical Oceanography, 2022, 41(4): 172-180.

图/表 8

图1

表1

表2

图2

表3

图3

表4

表5

参考文献 37

[1]	陈宝红, 谢尔艺, 高亚辉, 等, 2015. 米氏凯伦藻对海洋生物致毒作用的研究进展[J]. 福建水产, 37(3): 241-249.
	CHEN BAOHONG, XIE ERYI, GAO YAHUI, et al, 2015. Toxic effects of red tide caused by Karenia mikimotoi on marine organisms[J]. Journal of Fujian Fisheries, 37(3): 241-249. (in Chinese with English abstract)
[2]	陈旭阳, 刘保良, 2018. 海洋在线监测浮标在赤潮监测中的应用研究[J]. 热带海洋学报, 37(5): 20-24.
	CHEN XUYANG, LIU BAOLIANG, 2018. Application of Real-time monitoring buoy in monitoring red tide[J]. Journal of Tropical Oceanography, 37(5): 20-24. (in Chinese with English abstract)
[3]	赫芬·I·里斯. 2020. 机器学习实战: 使用R、tidyverse和mlr[M]. 但波, 高山, 韩建立, 译. 北京: 清华大学出版社: 1-536.
	RHYS H I. 2020. Machine learning: with R, the tidyverse, and mlr[M]. DAN BO, GAO SHAN, HAN JIANLI, trans. Beijing: Tsinghua University Publishing House: 1-536. (in Chinese)
[4]	蓝虹, 卢君峰, 曹宇峰, 等, 2014. 浅析平潭岛主要海洋灾害及防治对策[J]. 海洋开发与管理, 31(2): 55-58. (in Chinese)
[5]	李冬梅, 高永利, 田甜, 等, 2010. 水体扰动对多种赤潮藻生长的影响[J]. 热带海洋学报, 29(6): 65-70.
	LI DONGMEI, GAO YONGLI, et al, 2010. Effects of turbulence on phytoplankton: species differences[J]. Journal of Tropical Oceanography, 29(6): 65-70. (in Chinese with English abstract)
[6]	林森杰, 姬南京, 罗昊, 2019. 海洋有害藻华研究进展[J]. 海洋与湖沼, 50(3): 495-510.
	LIN SENJIE, JI NANJING, LUO HAO, 2019. Recent progress in marine harmful algal bloom research[J]. Oceanologia et Limnologia Sinica, 50(3): 495-510. (in Chinese with English abstract)
[7]	洛昊, 马明辉, 梁斌, 等, 2013. 中国近海赤潮基本特征与减灾对策[J]. 海洋通报, 32(5): 595-600.
	LUO HAO, MA MINGHUI, LIANG BIN, et al, 2013. Basic characteristics and mitigation countermeasures of red tides in China seas[J]. Marine Science Bulletin, 32(5): 595-600. (in Chinese with English abstract)
[8]	马毅, 吴瑞贞, 李华建, 等, 2008. 有利于赤潮消亡的水文气象条件[J]. 海洋预报, 25(3): 1-6.
	MA YI, WU RUIZHEN, LI HUAJIAN, et al, 2008. The hydrological-meteorological conditions conducive to the attenuation of red tide[J]. Marine Forecasts, 25(3): 1-6. (in Chinese with English abstract)
[9]	芮政, 杨桂军, 刘玉, 等, 2019. 扰动方式对水华微囊藻(Microcystis flos-aquae)群体大小的影响[J]. 湖泊科学, 31(2): 355-364. doi: 10.18307/2019.0205
	RUI ZHENG, YANG GUIJUN, LIU YU, et al, 2019. Effects of disturbance modes on the size of Microcystis flos-aquae colonies[J]. Journal of Lake Sciences, 31(2): 355-364. (in Chinese with English abstract) doi: 10.18307/2019.0205
[10]	苏金洙, 高佳, 苏玉萍, 等, 2020. 福建平潭海域米氏凯伦藻(Karenia mikimotoi)增殖影响因子研究[J]. 福建师范大学学报(自然科学版), 36(4): 43-49, 56.
	SU JINZHU, GAO JIA, SU YUPING, et al, 2020. Study on the factors affecting the proliferation of Karenia mikimotoi in Pingtan coastal area of Fujian province[J]. Journal of Fujian Normal University (Natural Science Edition), 36(4): 43-49, 56. (in Chinese with English abstract)
[11]	苏新红, 金丰军, 杨奇志, 等, 2017. 基于BP神经网络模型的福建海域赤潮预报方法研究[J]. 水产学报, 41(11): 1744-1755. (in Chinese)
[12]	小沃尔特·J·韦伯, 1980. 水质控制物理化学方法[M]. 上海市政工程设计院, 译. 北京: 中国建筑工业出版社: 510-511.
	WEBER W J, 1980. Physicochemical processes for water quality control[M]. Shanghai Municipal Engineering Design Institute, trans. Beijing: China Architecture & Building Press: 510-511. (in Chinese)
[13]	许昆灿, 暨卫东, 周秋麟, 等, 2004. 表观增氧量在近岸海域赤潮快速评价与预警中的应用[J]. 台湾海峡, 23(4): 417-422.
	XU KUNCAN, JI WEIDONG, ZHOU QIULIN, et al, 2004. Apparent oxygen increase (AOI) and its application in rapid assessment and prewarning of red tide in coastal waters[J]. Journal of Oceanography in Taiwan Strait, 23(4): 417-422. (in Chinese with English abstract)
[14]	许阳春, 张明峰, 苏玉萍, 等, 2020. 基于BP人工神经网络平潭海域赤潮叶绿素a浓度模型演算研究[J]. 海洋科学, 44(3): 34-41.
	XU YANGCHUN, ZHANG MINGFENG, SU YUPING, et al, 2020. Calculation of the Chlorophyll-a concentration of red tide in the Pingtan Coastal Zone by a BP artificial neural network model[J]. Marine Sciences, 44(3): 34-41. (in Chinese with English abstract)
[15]	俞志明, 陈楠生, 2019. 国内外赤潮的发展趋势与研究热点[J]. 海洋与湖沼, 50(3): 474-486.
	YU ZHIMING, CHENG NANSHENG, 2019. Emerging trends in red tide and major research progresses[J]. Oceanologia et Limnologia Sinica, 50(3): 474-486. (in Chinese with English abstract)
[16]	张海涵, 王娜, 宗容容, 等, 2022. 水动力条件对藻类生理生态学影响的研究进展[J]. 环境科学研究, 35(1): 181-190.
	ZHANG HAIHAN, WANG NA, ZONG RONGRONG, et al, 2022. Research progress on influence of hydrodynamic conditions on algal physiology and ecology[J]. Research of Environmental Sciences, 35(1): 181-190. (in Chinese with English abstract)
[17]	赵聪蛟, 刘希真, 付声景, 等, 2020. 基于水质浮标在线监测的米氏凯伦藻赤潮过程及环境因子变化特征分析[J]. 热带海洋学报, 39(2): 88-97.
	ZHAO CONGJIAO, LIU XIZHEN, FU SHENGJING, et al, 2020. Variation characteristics of the evolution of Karenia mikimotoi bloom and environmental factors based on online monitoring buoy data[J]. Journal of Tropical Oceanography, 39(2): 88-97. (in Chinese with English abstract)
[18]	赵雪, 杨凡, 郭娜, 等, 2009. 2007年2月汕头赤潮事件水文气象及海水理化因子影响分析[J]. 海洋预报, 26(1): 43-51. (in Chinese)
[19]	ANDERSON D M, 2009. Approaches to monitoring, control and management of harmful algal blooms (HABs)[J]. Ocean & Coastal Management, 52(7): 342-347.
[20]	AZANZA R V, FUKUYO Y, YAP L G, et al, 2005. Prorocentrum minimum bloom and its possible link to a massive fish kill in Bolinao, Pangasinan, Northern Philippines[J]. Harmful Algae, 4(3): 519-524. doi: 10.1016/j.hal.2004.08.006
[21]	BREIMAN L, 2001. Random forests[J]. Machine Learning, 45(1): 5-32. doi: 10.1023/A:1010933404324
[22]	DENG LEI, YANG WENYI, LIU HUI, 2019. PredPRBA: Prediction of protein-RNA binding affinity using gradient boosted regression trees[J]. Frontiers in Genetics, 10: 637. doi: 10.3389/fgene.2019.00637
[23]	ELITH J, LEATHWICK J R, HASTIE T, 2008. A working guide to boosted regression trees[J]. Journal of Animal Ecology, 77(4): 802-813. doi: 10.1111/j.1365-2656.2008.01390.x
[24]	JIMÉNEZ-CARVELO A M, GONZÁLEZ-CASADO A, BAGUR- GONZÁLEZ M G, et al, 2019. Alternative data mining/ machine learning methods for the analytical evaluation of food quality and authenticity - A review[J]. Food Research International, 122: 25-39. doi: 10.1016/j.foodres.2019.03.063
[25]	KARKI S, SULTAN M, ELKADIRI R, et al, 2018. Mapping and forecasting onsets of harmful algal blooms using MODIS data over coastal waters surrounding Charlotte County, Florida[J]. Remote Sensing, 10(10): 1656. doi: 10.3390/rs10101656
[26]	LIN H T, LIANG T J, CHEN S M, 2013. Estimation of battery state of health using probabilistic neural network[J]. IEEE Transactions on Industrial Informatics, 9(2): 679-685. doi: 10.1109/TII.2012.2222650
[27]	LIU RUONAN, YANG BOYUAN, ZIO E, et al, 2018. Artificial intelligence for fault diagnosis of rotating machinery: A review[J]. Mechanical Systems and Signal Processing, 108: 33-47. doi: 10.1016/j.ymssp.2018.02.016
[28]	LYU JUNCHENG, ZHANG JIE, 2019. BP neural network prediction model for suicide attempt among Chinese rural residents[J]. Journal of Affective Disorders, 246: 465-473. doi: 10.1016/j.jad.2018.12.111
[29]	NABIPOUR M, NAYYERI P, JABANI H, et al, 2020. Deep learning for stock market prediction[J]. Entropy, 22(8): 840. doi: 10.3390/e22080840
[30]	PRASAD A M, IVERSON L R, LIAW A, 2006. Newer classification and regression tree techniques: Bagging and random forests for ecological prediction[J]. Ecosystems, 9(2): 181-199. doi: 10.1007/s10021-005-0054-1
[31]	QIN MENGJIAO, LI ZHIHANG, DU ZHENHONG, 2017. Red tide time series forecasting by combining ARIMA and deep belief network[J]. Knowledge-Based Systems, 125: 39-52. doi: 10.1016/j.knosys.2017.03.027
[32]	RECKNAGEL F, 1997. ANNA-Artificial Neural Network model for predicting species abundance and succession of blue-green algae[J]. Hydrobiologia, 349(1): 47-57. doi: 10.1023/A:1003041427672
[33]	SAN DIEGO-MCGLONE M L, AZANZA R V, VILLANOY C L, et al, 2008. Eutrophic waters, algal bloom and fish kill in fish farming areas in Bolinao, Pangasinan, Philippines[J]. Marine Pollution Bulletin, 57(6-12): 295-301. doi: 10.1016/j.marpolbul.2008.03.028
[34]	TIAN WENCHONG, LIAO ZHENLIANG, ZHANG JIN, 2017. An optimization of artificial neural network model for predicting chlorophyll dynamics[J]. Ecological Modelling, 364: 42-52. doi: 10.1016/j.ecolmodel.2017.09.013
[35]	WU DANNI, CHEN JUNHUI, HE XIUPING, et al, 2019. Distribution, partitioning, and seasonal variation of lipophilic marine algal toxins in aquatic environments of a typical semi-closed mariculture bay[J]. Environmental Pollution, 255: 113299. doi: 10.1016/j.envpol.2019.113299
[36]	ZHOU BINGYIN, LU MING, WANG YONGGANG, 2016. Counting people using gradient boosted trees[C]// 2016 IEEE Information Technology, Networking, Electronic and Automation Control Conference. Chongqing, China: IEEE: 391-395.
[37]	ZOHDI E, ABBASPOUR M, 2019. Harmful algal blooms (red tide): a review of causes, impacts and approaches to monitoring and prediction[J]. International Journal of Environmental Science and Technology, 16(3): 1789-1806. doi: 10.1007/s13762-018-2108-x

数据类型	环境指标	数据来源
水质数据	水温、盐度、溶解氧、pH、氨氮、硝酸盐-氮、亚硝酸盐-氮、磷酸盐、硅酸盐、溶解氧饱和度、化学耗氧量、叶绿素a、浮游植物密度、气温共14个指标	福建省海洋环境与渔业资源监测中心
气象数据	风速、降水量、日照时数、气温共4个指标	中国气象数据网

模型	优点	缺点
KNN回归模型	理论成熟, 易于实现(Jiménez-Carvelo et al, 2019), 可以同时用于分类和回归(Liu et al, 2018)	需要大量运算以及很大的存储空间, 最邻近数据的选择对右方的影响太大(Liu et al, 2018; Jiménez-Carvelo et al, 2019)
RF回归模型	不会过度拟合数据, 随机预测器选择保持低偏差。适合不稳定模型或分类不平衡问题, 可提供更好的预测模型(Prasad et al, 2006; Jiménez-Carvelo et al, 2019)	在时间序列和计算机资源方面要求非常高。模拟结果复杂, 分类显示为非图形树(Prasad et al, 2006; Jiménez-Carvelo et al, 2019)
GBRT回归模型	能够处理不同类型的数据, 无需预先对数据进行转换或排除异常值, 可以适应复杂的非线性关系, 并自动处理预测器之间的交互效应(Deng et al, 2019)	会出现过度拟合, 因此可能无法拟合新样本, 从而无法正确预测(Jiménez-Carvelo et al, 2019)
Bagging回归模型	在降低高维数据集的方差和误差方面非常有效。训练集中没有使用的数据称为“out-of-bag”数据, 可以用来提供更好的误差估计(Prasad et al, 2006)	因为大量的回归树(通常是30~50棵)是平均的, 所以解释结果并不容易, 此外误差的偏倚分量略优于单回归树(Prasad et al, 2006)

	水温	盐度	溶解氧	AOI	pH	叶绿素a	风速	降水	日照	气温	海平面气压
盐度	0.298**
溶解氧	-0.707**	-0.401**
AOI	-0.147**	0.028	0.783**
pH	0.006	0.070*	0.152**	0.272**
叶绿素a	-0.039	-0.066	0.348**	0.479**	0.230**
风速	-0.011	0.170**	-0.174**	-0.057	-0.011	-0.134**
降水	-0.117**	-0.214**	0.066	-0.131**	-0.205**	-0.093**	-0.022
日照	0.351**	0.190**	-0.203**	0.003	0.031	-0.007	-0.149**	-0.332**
气温	0.844**	0.379**	-0.621**	-0.081	0.073*	0.034	-0.031	-0.281**	0.532**
海平面气压	-0.567**	-0.231**	0.443**	0.126**	-0.156**	-0.032	0.066	0.058	-0.099**	-0.686**
藻密度	0.057	0.076*	0.178**	0.251**	0.107**	0.414**	-0.109**	0.006	-0.013	0.112**	-0.123**

组合	指标	KNN回归模型			RF回归模型			GBRT回归模型			Bagging回归模型
组合	指标	R²	MAE/ (μg·L^-1)	RMSE/ (μg·L^-1)	R²	MAE/ (μg·L^-1)	RMSE/ (μg·L^-1)	R²	MAE/ (μg·L^-1)	RMSE/ (μg·L^-1)	R²	MAE/ (μg·L^-1)	RMSE/ (μg·L^-1)
组合1	风速+气温+海平面气压+叶绿素a	0.624	0.836	0.821	0.596	0.836	0.829	0.562	0.829	0.839	0.585	0.832	0.839
组合2	气温+海平面气压+叶绿a+AOI	0.363	0.814	0.896	0.468	0.825	0.865	0.485	0.824	0.862	0.496	0.822	0.867
组合3	风速+气温+海平面气压+叶绿a+AOI	0.477	0.823	0.863	0.537	0.829	0.846	0.536	0.826	0.846	0.530	0.826	0.853

输入组合	输入因子	RMSE/(μg·L^-1)	MAE/(μg·L^-1)	R²
组合1	气温+日照+溶解氧	0.066	0.033	0.499
组合2	气温+日照+AOI	0.073	0.041	0.258
组合3	气温+日照+风速	0.048	0.033	0.333
组合4	气温+日照+降水+溶解氧	0.053	0.031	0.602
组合5	气温+日照+风速+AOI	0.062	0.033	0.651
组合6	气温+日照+风速+溶解氧	0.046	0.027	0.290

福建平潭近海赤潮预警模型研究^*

Study on the early warning model of red tide in the offshore area of Pingtan, Fujian province

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 37

相关文章 13

Metrics

本文评价

推荐阅读 0

[1]	王朝晖, 张宇宁, 王文婷, 谢昌良, 陈佳卓, 郑虎, 王军星. 福建东山湾表层沉积物中甲藻孢囊分布研究[J]. 热带海洋学报, 2022, 41(4): 154-162.
[2]	李傲, 冯洋, 王云涛, 薛惠洁. 基于OC-CCI数据的南海高叶绿素a浓度水域面积的时空变化研究*[J]. 热带海洋学报, 2022, 41(2): 77-89.
[3]	王会芳, 黄秀清, 刘建华, 徐美娜, 蒋芸芸, 邱桔斐. 福建海坛海峡赤潮灾害潜在生态风险评估[J]. 热带海洋学报, 2021, 40(4): 122-133.
[4]	潘翠红, 夏丽华, 吴志峰, 王猛, 解学通, 王芳. 柘林湾近岸水产养殖区水域叶绿素a浓度反演[J]. 热带海洋学报, 2021, 40(1): 142-153.
[5]	张雨, 王桂芬, 许占堂, 杨跃忠, 周雯, 郑文迪, 曾凯, 邓霖. 基于浮标观测的东海赤潮高发区秋、冬季漫射衰减系数的反演[J]. 热带海洋学报, 2020, 39(5): 71-83.
[6]	赵聪蛟, 刘希真, 付声景, 姚炜民, 周燕, 马骏. 基于水质浮标在线监测的米氏凯伦藻赤潮过程及环境因子变化特征分析[J]. 热带海洋学报, 2020, 39(2): 88-97.
[7]	陈旭阳, 刘保良. 海洋在线监测浮标在赤潮监测中的应用研究[J]. 热带海洋学报, 2018, 37(5): 20-24.
[8]	杨超宇, 唐丹玲, 叶海彬. 基于GF-4遥感数据的叶绿素浓度反演算法研究^*[J]. 热带海洋学报, 2017, 36(5): 33-39.
[9]	傅丽君, 林天城, 余晓琪, 郑天凌. 抑藻细菌HSY-03的分离、鉴定及对赤潮异弯藻生长的影响[J]. 热带海洋学报, 2016, 35(3): 87-93.
[10]	王青 , 朱良生, 胡金鹏. 最优插值法在东海赤潮高发区赤潮生消过程数值模拟中的应用[J]. 热带海洋学报, 2012, 31(6): 29-35.
[11]	尹平河,黄凤,赵玲. 载Fe3+纳米TiO2薄膜去除球形棕囊藻赤潮生物的研究[J]. 热带海洋学报, 2010, 29(4): 102-106.
[12]	曹文熙,杨跃忠,王桂芬,赵俊,周雯,李彩,孙兆华,梁少君,柯天存,卢桂新. 赤潮的高光谱监测及预警方法[J]. 热带海洋学报, 2010, 29(2): 17-24.
[13]	欧林坚,张玉宇,李扬,王华接,谢学东,荣志明,吕颂辉,齐雨藻,. 广东珠海双胞旋沟藻Cochlodinium geminatum赤潮事件分析[J]. 热带海洋学报, 2010, 29(1): 57-61.